All of natural science, by definition, has the primary task of reducing Nature's complexities given to a limited set of concepts, principles, laws, and mechanisms that, more or less, captures that complexity. Behavior analysis, as a natural science, certainly exemplifies this task, by, for example, reducing the complexities of behavior to the action of contingencies. But even within that framework, there remain many complexities. I will discuss some of the "homogeneous" modes and issues of reduction in behavior analysis, but there are even greater challenges. As essentially a biological science, the issue arises of the relationships of behavior analysis to other biological functions at all levels-especially behavior in relation to neural processes-so called "heterogeneous" reduction. This is one of the most difficult issues in biological science. The challenges are not simply empirical, but deeply conceptual. I will attempt to survey some of these challenges.